Stimulation and inhibition of human mammary epithelial cell duct morphogenesis in vitro.

We previously determined that normal human mammary epithelial cells (HMECs) placed on the basement membrane-like substance Matrigel form structures, whereas malignant breast cells do not (1). In the present study, we determined that the structures formed by normal cells on Matrigel resembled breast ducts in vivo by electron microscopy, and the process of their formation recapitulated what is known of duct formation in vivo. We therefore used this model to study less well-understood aspects of breast morphogenesis. Two priming signals appeared necessary for initiation of morphogenesis: one provided by the Matrigel and one by the cells in an autocrine fashion. Evidence for this included diminished duct formation by cells plated low-concentration Matrigel or at low cell densities, and the reversal of the latter by conditioned medium from high-density cells on Matrigel. Antibodies to bFGF inhibited morphogenesis, suggesting a stimulatory autocrine role for this factor, and antibodies to TGF-beta 1 stimulated duct formation, suggesting an inhibitory autocrine role. Added TGF-beta 1 abolished morphogenesis and stimulated normal cells to wander through Matrigel as do malignant cells. Conditioned medium from normal cells did not stimulate malignant cells to form ducts, but conditioned medium from tumor cells diminished normal morphogenesis, suggesting that malignant cells secrete an inhibitor of morphogenesis.

Expression of hemidesmosomes and component proteins is lost by invasive breast cancer cells.

Hemidesmosomes are multiprotein structures that attach basal cells of stratified epithelia to basement membranes. Although normal human breast epithelia are not stratified, we observed expression of electron-dense hemidesmosomes and hemidesmosome protein components by breast epithelial and myoepithelial cells at the basal lamina in vivo. Primary cultured normal human breast epithelial cells also contained hemidesmosomes and component proteins, and could be used as a model for hemidesmosome assembly and regulation. In these cultured cells, hemidesmosome proteins were expressed and localized basally in an unvaried temporal pattern, and electron-dense hemidesmosomes were not seen until the final protein was localized to the cell base. In addition, rate of localization was influenced by confluence, doubling time, and extracellular matrix. Invasive breast cancer cells did not express hemidesmosomes or most of the component proteins in vivo. In carcinoma in situ, cells away from the basement membrane lacked hemidesmosomes and hemidesmosome proteins, and cells at the basement membrane exhibited abnormalities of hemidesmosome protein expression. Primary human malignant breast cells in culture exhibited a mix of hemidesmosome phenotypes. These data suggest that hemidesmosomes may be important subcellular structures in determining the cytoarchitecture of the breast epithelium. Further, their downregulation may influence cytoarchitecture remodeling closely linked with cell cycle, motility, and extracellular matrix interactions; and their loss in carcinoma may be associated with loss of normal cytoarchitecture.

Alterations in integrin and basement-membrane protein expression by malignant breast cells.

We investigated expression of eight integrin subunits and four basement membrane proteins by normal and malignant human mammary epithelial cells (HMEC) in culture and in tissue sections. Normal HMEC expressed the same six integrin subunits in culture and in vivo. Expression of some of these was altered in vivo in breast carcinomas: alpha1 was not observed and beta4 was missing from some cells-changes that were not genetic, as malignant cells in primary culture could express these subunits. Nor was expression of these subunits related to expression of their basement membrane protein ligands. Other integrin subunits were reduced in culture, and expression of basement membrane proteins differed between normal and malignant cells and cells in vivo and in culture. Our results suggest that breast carcinoma cells have altered regulation of expression of adhesion receptors and their ligands, and these alterations can be manipulated in culture for further study.

Culture of normal and malignant primary human mammary epithelial cells in a physiological manner simulates in vivo growth patterns and allows discrimination of cell type.

We cultured primary human mammary epithelial cells from five reduction mammoplasties and five breast carcinomas and attempted to improve culture conditions and define cell populations grown. Normal cells cultured on Matrigel basement membrane-like substance formed multicellular three-dimensional structures reminiscent of tissue ducts and alveoli, while malignant cells remained as single cells crawling through Matrigel much as malignant cells separate and invade basement membrane in vivo. This re-creation of normal and malignant breast cell morphology may facilitate studies of breast cancer cell biology and determination of malignant cell authenticity in culture. Growth of cells in a reduced oxygen concentration of 12% improved cell proliferation over room air (21%); however, cells could not proliferate in a completely physiological oxygen concentration of 6%, perhaps because of the medium used. We developed an improved medium for malignant cell growth, which lengthened their life span in culture, and a completely defined medium which supported cell proliferation for six passages. Methods to determine the epithelial nature of mammary epithelial cells are illustrated and discussed. The authenticity of malignant cells in culture was suggested by their proliferation without certain growth factors required for normal cell growth or with transforming growth factor-beta, which arrests normal cell proliferation, and by their contact independence.

Comparison of fatty acid and triacylglycerol metabolism of macrophages and smooth muscle cells.

The response of macrophages and smooth muscle cells to culture in free fatty acid has been compared. Because oleate and linoleate promoted triacylglycerol enrichment of smooth muscle cells, whereas palmitate had little effect, oleate was used for these studies. The kinetics of the accumulation of triacylglycerol produced by oleate was comparable between smooth muscle cells and macrophages. When grown in increasing concentrations of oleic acid at various fatty acid to albumin molar ratios, the extent of triacylglycerol accumulation in both cell types was dependent on the concentration of oleate, the concentration of albumin, and the oleate to albumin molar ratio. However, macrophages contained 2.6-fold more triacylglycerol than smooth muscle cells in the presence of oleate at 0.36 mM or greater and at levels of albumin higher than 0.15 mM. The cellular triacylglycerol content of macrophages was linearly related to the oleate to albumin molar ratio at both a constant albumin concentration and a constant oleate concentration, whereas the accumulation of triacylglycerol in smooth muscle cells showed a curvilinear relationship. When cells were preloaded with triacylglycerol, smooth muscle cells showed a greater loss of lipid when exposed to albumin than macrophages did. Over a two-hr time period, macrophages incorporated twice as much labeled fatty acid as smooth muscle cells. Thus, while smooth muscle cells and macrophages showed similar responses to exogenous fatty acid and albumin, there were also significant quantitative distinctions.

Modification of very low density lipoproteins leads to macrophage scavenger receptor uptake and cholesteryl ester deposition.

Chemically modified low density lipoproteins (LDL) are recognized by the macrophage scavenger receptor and can lead to substantial cholesteryl ester accumulation in cultured macrophages. Uptake of modified lipoproteins in vivo could contribute to foam cell formation during generation of the atherosclerotic plaque lesion. In the present study, modification of human pre-beta migrating very low density lipoprotein (VLDL) by acetylation led to recognition by the macrophage scavenger receptor as demonstrated in cross-competition experiments with acetylated LDL (ALDL). Recognition by this alternative binding site was associated with increased cholesterol delivery to human macrophages as assessed by suppression of LDL receptor activity, stimulation of cholesterol esterification rates, and accumulation of intracellular cholesteryl ester. Subfractionation of acetylated very low density lipoprotein (AVLDL) by ultracentrifugation in a discontinuous NaCl gradient demonstrated that AVLDL subfractions were equally effective in competing for 125I-ALDL uptake by macrophages when compared on the basis of particle number. These results suggest that modification of VLDL with subsequent recognition by the macrophage scavenger receptor may be a mechanism by which VLDL particles participate in macrophage cholesteryl ester overload.

Macrophage interaction with very-low-density lipoproteins results in triacylglycerol-enriched smooth muscle cells.

Macrophage-conditioned medium containing very-low-density lipoproteins (VLDL) and its effects on smooth muscle cell triacylglycerol metabolism was investigated. Macrophages exposed to VLDL from normolipemic rats accumulated high levels of intracellular triacylglycerol, while similarly treated smooth muscle cells showed only slight changes. Medium, initially composed of VLDL and albumin, contained substantial levels of free fatty acids after exposure to macrophages. In the presence of albumin, the loss of VLDL triacylglycerol from the medium and the appearance of medium free fatty acids was consistent with a precursor-product relationship. The extent of medium fatty acid accumulation was dependent on the length of time of incubation with macrophages and was proportional to the concentration of VLDL and albumin added to the culture dish. This macrophage-conditioned medium, when given to smooth muscle cells, promoted a 6-12-fold increase in smooth muscle cell triacylglycerol levels over that produced by fresh VLDL and albumin. Similar increases in cell triacylglycerol levels could be produced using fresh medium approximating the oleate concentration and the fatty acid to albumin molar ratios found in macrophage-conditioned medium. In macrophage-conditioned medium with VLDL but without albumin, little free fatty acid was found. Other factors produced by macrophages did not seem to affect the metabolism of VLDL by smooth muscle cells since, in the absence of albumin, media with VLDL caused comparable responses in smooth muscle cell triacylglycerol accumulation whether or not the medium was previously exposed to macrophages. Thus, the minor changes in triacylglycerol content in smooth muscle cells promoted by medium containing VLDL and albumin were substantially enhanced by a prior exposure of the medium to macrophages, primarily due to the free fatty acids present in the macrophage-conditioned medium.